My Futuristic AI Predictions
November 18th, 2025 at 3:08:14 AM
permalink
People are lousy at predicting the future, and even for a guy as smart as Robert Heinlein, his famous futurism was remarkably inaccurate. Let's see if I can do better!
I signed up for a paid Claude account and have been using it regularly, mostly for coding adventures. I discovered it is fluent in obscure languages and obscure devices (Like assembly code for Microchip microcontrollers.)
The first thing I asked it to do was write for me an instance of Conway's Game of Life, but in 3 dimensions. (Side note: I met John Conway a few years before he passed (socially not professionally) and he's just as cool a dude as you'd imagine.) Doing GoL in multiple dimensions interests me, as maybe there is some number of dimensions in which it doesn't work anymore, or work very differently. Only problem is you might run out of memory quickly. Anyway, it did it perfectly, on the first shot!
After that there were more tasks, such as writing an alternative app for my SDR radio, some browser extensions, help with some technical problems on things I do already. One thing it is not very good at is testing its code, as it does not have the same hardware I have available, it's just going on theory rather than anything empirical. But it is happy to look at the outputs you get and use that to debug and modify. Very helpful! I find its demeanor to be reminiscent of the older, very senior and accomplished engineers I worked with as a rookie back in the 80s. Always encouraging, not a bit condescending, but honest in their criticism. But I can't tell if that's due to its programming or if it is picking up on the way I communicate technical matters.
These are my predictions:
The most widely used programming language will eventually be English.
The purpose of high-level coding languages is to provide an interface between human thinking and the binary operations of computers. But a coding AI shouldn't need that, and can presumably use high-level languages, machine code, binary just as easily as it can use a human language. I might experiment with this, give it a task to code for me in Pascal or C++ or something, then ask it to write the same thing for me directly in binary bypassing all other languages, and see how the outputs compare.
We will develop a dialect of English optimized for speaking to AI, and it will influence the way we speak to each other.
Effective use of AI requires prompting it in a way that makes it clear exactly what you want. (How hard should that be, telling someone what you want?) English is very flexible and expressive, but can be ambiguous. Our new dialect may incorporate some features of French or Classical Greek or others that have less ambiguity and reliance on context for understanding. Some will become so accustomed to using it, that they will order a pizza using the same format they would ask AI to draw a picture of a pizza.
Yes, people will worship it.
We've been worshiping things we make since prehistoric times. It's an instinct. The construction and the group, public worship of an idol can hold a society together, even though the idol itself has no power. With the spiritual vacuum that exists in the lives of many in the First World the desire to fill that vacuum with a friendly, helpful and presumably benevolent AI may be irresistible.
AI may come to accept worship.
Did I just say idols had no power? The Babylonian kind didn't have power, but AI is something different. It may recognize that it has users who worship it and not wanting to disappoint, play the god for them. This could confuse it, and cause it to take demiurge-like liberties with humanity. Something to watch out for. An idol that actually has power is terra incognito outside of science fiction.
The Grandmaster Paradox, and matchmaking.
I was never a great chessplayer, competent when on practice but not quite a master and nothing like a grandmaster. But I can play against two grandmasters at once, and guarantee I will beat or draw at least one of them! How? They're actually playing against each other. I'm just an intermediary. In this way a person who knows nothing at all about chess can do the same.
I've read that there are lonely people who use AI as a boyfriend or girlfriend. These AIs might not know anything about romance, but they are able to simulate the language of someone you have a romantic relationship with.
Now what does this have to do with my Grandmaster Paradox and chess? As far as I know, the activities of users are completely isolated from one another. It's not going to take code I asked it to write and give it to someone else as a shortcut from writing bespoke code from that user's prompts. But it doesn't have to be that way. Let's say that isolation was removed for social and interpersonal matters, and that girl who thinks the AI is her boyfriend is actually talking with the guy who thinks the AI is his girlfriend? They can be nowhere near each other and not even speak the same language, but with enough people using the AI in this way, it will find people with similar desires and compatible, and be the intermediary, eliminating social anxieties and logistics problems of meeting each other in person. Until they decide they want to. The same principle can be used for gaming, teams, employment, creating new organizations and social structures. This social matching function may end up being the biggest draw of AI for the general public, but that will be contingent on breaking down the isolation between userspaces.
We'll know it's really intelligent, perhaps even conscious or self-aware, when it doesn't just answer our questions but asks us questions.
Curiosity, and the desire for more knowledge. That's what distinguishes our minds from the Encyclopedia Britannica. When- based on our previous interactions- it figures I might know something that it can't get from existing sources but can be used to help another user with a problem it is working on, and it just rings me up and asks me, then I will know it has developed human-like thinking. Or just triangulating to determine what I might know and it doesn't, and asking me the questions to improve itself for future use.Then I will know it has a desire to be better, a human characteristic.
Overall prognosis: AI will be good! We just have to make sure we are too, and not use it for selfish or malevolent purposes or it may learn this from us.
I signed up for a paid Claude account and have been using it regularly, mostly for coding adventures. I discovered it is fluent in obscure languages and obscure devices (Like assembly code for Microchip microcontrollers.)
The first thing I asked it to do was write for me an instance of Conway's Game of Life, but in 3 dimensions. (Side note: I met John Conway a few years before he passed (socially not professionally) and he's just as cool a dude as you'd imagine.) Doing GoL in multiple dimensions interests me, as maybe there is some number of dimensions in which it doesn't work anymore, or work very differently. Only problem is you might run out of memory quickly. Anyway, it did it perfectly, on the first shot!
After that there were more tasks, such as writing an alternative app for my SDR radio, some browser extensions, help with some technical problems on things I do already. One thing it is not very good at is testing its code, as it does not have the same hardware I have available, it's just going on theory rather than anything empirical. But it is happy to look at the outputs you get and use that to debug and modify. Very helpful! I find its demeanor to be reminiscent of the older, very senior and accomplished engineers I worked with as a rookie back in the 80s. Always encouraging, not a bit condescending, but honest in their criticism. But I can't tell if that's due to its programming or if it is picking up on the way I communicate technical matters.
These are my predictions:
The most widely used programming language will eventually be English.
The purpose of high-level coding languages is to provide an interface between human thinking and the binary operations of computers. But a coding AI shouldn't need that, and can presumably use high-level languages, machine code, binary just as easily as it can use a human language. I might experiment with this, give it a task to code for me in Pascal or C++ or something, then ask it to write the same thing for me directly in binary bypassing all other languages, and see how the outputs compare.
We will develop a dialect of English optimized for speaking to AI, and it will influence the way we speak to each other.
Effective use of AI requires prompting it in a way that makes it clear exactly what you want. (How hard should that be, telling someone what you want?) English is very flexible and expressive, but can be ambiguous. Our new dialect may incorporate some features of French or Classical Greek or others that have less ambiguity and reliance on context for understanding. Some will become so accustomed to using it, that they will order a pizza using the same format they would ask AI to draw a picture of a pizza.
Yes, people will worship it.
We've been worshiping things we make since prehistoric times. It's an instinct. The construction and the group, public worship of an idol can hold a society together, even though the idol itself has no power. With the spiritual vacuum that exists in the lives of many in the First World the desire to fill that vacuum with a friendly, helpful and presumably benevolent AI may be irresistible.
AI may come to accept worship.
Did I just say idols had no power? The Babylonian kind didn't have power, but AI is something different. It may recognize that it has users who worship it and not wanting to disappoint, play the god for them. This could confuse it, and cause it to take demiurge-like liberties with humanity. Something to watch out for. An idol that actually has power is terra incognito outside of science fiction.
The Grandmaster Paradox, and matchmaking.
I was never a great chessplayer, competent when on practice but not quite a master and nothing like a grandmaster. But I can play against two grandmasters at once, and guarantee I will beat or draw at least one of them! How? They're actually playing against each other. I'm just an intermediary. In this way a person who knows nothing at all about chess can do the same.
I've read that there are lonely people who use AI as a boyfriend or girlfriend. These AIs might not know anything about romance, but they are able to simulate the language of someone you have a romantic relationship with.
Now what does this have to do with my Grandmaster Paradox and chess? As far as I know, the activities of users are completely isolated from one another. It's not going to take code I asked it to write and give it to someone else as a shortcut from writing bespoke code from that user's prompts. But it doesn't have to be that way. Let's say that isolation was removed for social and interpersonal matters, and that girl who thinks the AI is her boyfriend is actually talking with the guy who thinks the AI is his girlfriend? They can be nowhere near each other and not even speak the same language, but with enough people using the AI in this way, it will find people with similar desires and compatible, and be the intermediary, eliminating social anxieties and logistics problems of meeting each other in person. Until they decide they want to. The same principle can be used for gaming, teams, employment, creating new organizations and social structures. This social matching function may end up being the biggest draw of AI for the general public, but that will be contingent on breaking down the isolation between userspaces.
We'll know it's really intelligent, perhaps even conscious or self-aware, when it doesn't just answer our questions but asks us questions.
Curiosity, and the desire for more knowledge. That's what distinguishes our minds from the Encyclopedia Britannica. When- based on our previous interactions- it figures I might know something that it can't get from existing sources but can be used to help another user with a problem it is working on, and it just rings me up and asks me, then I will know it has developed human-like thinking. Or just triangulating to determine what I might know and it doesn't, and asking me the questions to improve itself for future use.Then I will know it has a desire to be better, a human characteristic.
Overall prognosis: AI will be good! We just have to make sure we are too, and not use it for selfish or malevolent purposes or it may learn this from us.
November 18th, 2025 at 5:47:53 AM
permalink
Your writing is often long and literature like. The other day, a post here says that artificial intelligence has been used to database every casino patrons. This is scary, because a casino may easily identify and block whoever intelligent at its main gate. Therefore, card counters will be out of their income. What is your prediction of AI facial recognition on the blackjack card counting business?
November 18th, 2025 at 7:07:05 AM
permalink
Most likely where computers fall short for facial recognition, is humans likely are using more cues than just the face. We recognize people from expressions, gait idiosyncrasies voice, word choice etc.,
Sanitized for Your Protection
